Research on the Completeness of the Injury and Illness Counts from the Survey of Occupational Injuries and Illnesses

The Bureau of Labor Statistics (BLS) is conducting ongoing research to investigate the completeness of the injury and illness counts from the Survey of Occupational Injuries and Illnesses (SOII). The purpose of this research is to better understand the reasons for an undercount of occupational injuries and illnesses reported by the SOII and to identify ways to improve the collection of occupational injury and illness data.

This page will be updated as additional research is undertaken and results are published. If you have any questions regarding this ongoing research, please contact us at or (202) 691-6170.

History and Expansion of the SOII

Following passage of the Occupational Safety and Health Act of 1970 the BLS, in cooperation with participating state governments, designed and conducted the first SOII in 1972 in order to provide the nation with reliable annual estimates of the number and rate of workplace injuries and illness by detailed industry. Since this initial publication over 40 years ago, BLS has continually worked to provide a more comprehensive accounting of the injuries and illnesses that befall workers in the United States.

  • In 1992, SOII began publishing case and demographic data for injuries and illnesses to workers that resulted in at least one day away from work.
  • In 2008, SOII began publishing injury and illness data for workers in state and local governments.
  • In 2011 BLS began a pilot study involving the collection of limited case and demographic data for injuries and illnesses to workers that result in a job transfer or restriction, but no missed days away from work.

BLS takes outside criticism of its injury and illness data quality and completeness very seriously, and this scrutiny has often resulted in program changes and enhancements such as those noted above. For a more comprehensive look at the development of the SOII and the enhancements to BLS occupational injury and illness data collection since 1972, see The BLS Survey of Occupational Injuries and Illnesses: A Primer by current BLS Deputy Commissioner William J. Wiatrowski.

Research into the Completeness of the SOII Data

In the mid 2000’s, multiple media reports of unsafe workplaces along with several academic research papers matching SOII data to workers’ compensation records across several states pointed to a potential undercount of nonfatal occupational injuries and illnesses by the SOII. In 2007 and 2008 Congress held a series of hearings concerning workplace safety and health, generating additional interest from both researchers and policymakers on the issue of an undercount. In an article published in the August 2008 Monthly Labor Review, former BLS Assistant Commissioner for Safety, Health, and Working Conditions John W. Ruser examined the existing evidence for an undercount of occupational injuries and illnesses and highlighted initial steps taken by the BLS to address the issue. The following year, BLS established an ongoing research program to investigate the completeness of the SOII injury and illness counts.

BLS conducted a first round of research between 2009 and 2012, coordinating with several State Agency grantees and a contractor. Building on previous studies, BLS and its research partners matched SOII injury and illnesses cases to workers’ compensation records and the results indicated that SOII was not capturing all injuries and illnesses within its scope. However, the estimated magnitude of the undercount varied dramatically based on the case matching methodologies employed by the researchers in each State.

BLS and its research partners also attempted to enumerate the total number of all work-related amputations and instances of carpal tunnel syndrome, which included cases beyond the scope of the SOII, in three states. Researchers used multiple data sources, including the SOII, workers’ compensation claims, and hospital and emergency room records in this enumeration. As a result of the technical challenges, resource requirements, and varying availability of data sources across states in this study, BLS and its research partners determined that a multisource enumeration of workplace injuries and illnesses was not feasible at a national level but might be within a State.

Lastly, both BLS and state researchers conducted a small number of employer interviews with SOII respondents regarding their injury and illness recordkeeping procedures. While qualitative in nature, these interviews were valuable for understanding some of the SOII and workers’ compensation matching results and led to a much larger number of employer interviews across four states in the next round of research.

For additional discussion of the first round projects, see Examining the completeness of the occupational injury and illness data: an update on current research by William J. Wiatrowski, which appeared in the June 2014 edition of Monthly Labor Review. In addition, journal articles by BLS funded researchers on the three projects from Round 1 were published in the October 2014 issue of the American Journal of Industrial Medicine. Working paper versions of these articles can be found on the SOII Undercount Research Page.

In the second round of research (2012-2014) BLS worked with four states to conduct a large number of employer interviews with SOII respondents in each state about their injury and illness recordkeeping practices. Initial results from this study, as noted in the final state reports, point to confusion among some respondents about various recordkeeping rules that could result in injury and illness cases going unreported to BLS or the specific details about a case being erroneously recorded and reported. BLS has partnered with a research fellow to conduct a more detailed analysis of the results of these interviews, with a final report due later in 2016.

BLS also worked with Washington State on two additional projects in Round 2. The first involved a twelve year match (2000 – 2011) of SOII and workers compensation records to investigate any undercount trends over time. Washington found about 70 percent of eligible workers compensation time loss claims were found in SOII, implying a SOII undercount of about 30 percent. Washington also found a slight increase in SOII underreporting across that time period. More information is available in Washington’s final report to BLS and in their article in the April 2016 American Journal of Industrial Medicine.

For the second project, Washington State researchers interviewed a small number of in-state establishments in 2012 that filed what appeared to be SOII-eligible workers compensation claims but did not report the injury or illness on the SOII. This project was qualitative in nature as its goal was to identify as many different reasons as possible employers gave for not reporting these injuries or illnesses to BLS. More information is available in Washington’s final report to BLS and their article in the May 2016 American Journal of Industrial Medicine.

In late 2014 BLS began a third round of research, initiating two new projects. BLS worked with Westat to conduct a national follow-back survey of 2013 SOII respondents in order to collect nationwide data on employer injury and illness recordkeeping practices, including new information on the extent that injury and illness records are updated by employers in the year following an injury or illness. Data collection is complete and initial analysis of this data is underway. An overview of the data collection portion of this project can be found in the Westat final report to BLS.

BLS also worked with Westat to begin exploratory research on various approaches to collecting injury and illness data from workers that would complement the data BLS currently collects from employers via the SOII. Injury and illness data collected from workers via a household survey may help BLS better quantify the magnitude of the undercount of injuries and illnesses from the SOII. BLS used the recommendations provided by Westat in their final report to continue research on a worker survey of injuries and illnesses with the National Opinion Research Center (NORC) in 2015 and 2016. NORC has provided BLS guidance on the suitability of existing surveys and frames for a worker survey and also developed recommended survey designs. BLS is currently working with NORC to develop and test a worker survey questionnaire. BLS will use this questionnaire in a pilot worker survey planned for 2017-18.

BLS is also committed to expanding the use of automated coding techniques in SOII production processes to improve coding consistency and quality and is continuing research in this area that began in 2012. For an overview of manual and automated approaches to coding SOII data as well as a review of previous efforts to automate various coding tasks, see Automated Coding of Worker Injury Narratives by BLS economist Alexander Measure.

Publications authored/co-authored by BLS researchers

Publications from BLS funded State Partners and Contractors